ViDAQ: A computer vision based remote data acquisition system for reading multi-dial gauges

Abstract

Abstract This paper presents and evaluates design improvements to the Visual Data Acquisition (ViDAQ) system for reading multi-dial gauges. ViDAQ in general, is targeted to occupy a niche application for a cost effective and readily deployable solution for non-intrusive and remote acquisition of data from legacy human machine interface (HMI) indicators. Legacy HMI indicators that pose numerous technological hurdles in being digitally monitored, include analogue rotary multi-dial gauges, alarm lamps, switches etc, much like those common to industrial process monitoring systems can benefit from ViDAQ. Furthermore, ViDAQ is poised to assist in realizing an overarching design goal of a generic EYE-on-HMI (Expert supervisorY systEm) framework. As a framework, EYE-on-HMI stands to integrate the burgeoning field of machine learning and computer vision for real-time detection of human-in-the-loop operator errors and gather human performance data in any commercial and/or industrial process control domain for improving operational safety. Operator interaction with HMI is vital to the operational safety of any process control such as in nuclear power plant operation, aviation, public transit vehicles, driverless vehicles, etc. and thus should be monitored actively.